REVIEW URRENT C OPINION
Adaptive servo-ventilation for the treatment of central sleep apnea in congestive heart failure: what have we learned? Lee K. Brown a and Shahrokh Javaheri b
Purpose of review Positive airway pressure devices for the noninvasive treatment of sleep-disordered breathing are being marketed that have substantially expanded capabilities. Most recently, adaptive servo-ventilation devices have become available that are capable of measuring patient ventilation continuously and use that information to adjust expiratory positive airway pressure and pressure support levels to abolish central and obstructive apneas and hypopneas, including central sleep-disordered breathing of the Hunter–Cheyne– Stokes variety. Patients with congestive heart failure are particularly prone to developing central sleep apnea and/or Hunter–Cheyne–Stokes breathing, and studies have shown that suppression of these abnormal breathing patterns may improve cardiac function and, ultimately, mortality. Recent findings Over the last approximately 18 months, increasing numbers of studies have appeared demonstrating improvement in cardiac function and other important outcomes after both acute application of adaptive servo-ventilation as well as 3 to 6 months of use in patients with congestive heart failure and central sleep apnea/Hunter–Cheyne–Stokes breathing. Several of these studies are randomized controlled trials and several include assessment of cardiac event-free survival showing an advantage to treating with adaptive servo-ventilation. Summary As an adjunct to optimal pharmacological management, adaptive servo-ventilation shows considerable promise as a means to improve outcomes in patients with congestive heart failure complicated by central sleep apnea/Hunter–Cheyne–Stokes breathing. Larger randomized controlled trials will be necessary to demonstrate the ultimate role of this therapeutic modality in such patients. Keywords adaptive servo-ventilation, central sleep apnea, congestive heart failure, Hunter–Cheyne–Stokes breathing
INTRODUCTION The last decade has seen the introduction of evermore technologically advanced devices for the noninvasive treatment of sleep-disordered breathing (SDB). Upon the original, straightforward platform of the continuous positive airway pressure (CPAP) flow generator, first introduced in 1981, have been built devices with a dizzying array of capabilities. These include bilevel positive airway pressure (biPAP) flow generators that switch between a lower pressure during expiration [expiratory positive airway pressure (EPAP)] and a higher pressure during inspiration [inspiratory positive airway pressure (IPAP)]; automatically titrating flow generators that detect decreases in airflow and signs of upper airway obstruction in order to vary CPAP or biPAP settings; and devices that target either a set level of www.co-pulmonarymedicine.com
ventilation (average volume assured pressure support) or a proportion of the patient’s native ventilation or inspiratory flow [adaptive servoventilation (ASV)]. The latter devices now combine their original technology with auto-titrating EPAP. There are currently two manufacturers marketing ASV flow generators in the United States (ResMed, a Department of Internal Medicine, University of New Mexico School of Medicine, Albuquerque, New Mexico, bUniversity of Cincinnati College of Medicine, and Cincinnati, Ohio, USA
Correspondence to Lee K. Brown, MD, Program in Sleep Medicine, University of New Mexico Health Sciences Center, 1101 Medical Arts Avenue NE, Building 2, Albuquerque, NM 87102, USA. Tel: +1 505 272 6110; fax: +1 505 272 6112; e-mail: [email protected] Curr Opin Pulm Med 2014, 20:550–557 DOI:10.1097/MCP.0000000000000108 Volume 20 Number 6 November 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Adaptive servo-ventilation Brown and Javaheri
KEY POINTS ASV, when properly titrated, has proven efficacy in suppressing CSA/HCSB in patients with CHF. Recent investigations, including several RCTs of high quality, have shown improvement in cardiac function, exercise capacity, and event-free survival when CHF patients receive ASV therapy to suppress CSA/HCSB. Larger RCTs will be necessary to confirm the proper role of ASV as a component of CHF management.
San Diego, California markets the S9 VPAP Adapt, and Philips Respironics, Murrysville, Pennsylvania markets the biPAP autoSV Advanced – System One) and a third device is available outside the United ¨r Medizin GmbH þ Co., States (Weinmann Gera¨te fu Hamburg, Germany produces the SOMNOvent CR) [1 ]. All three devices have shown particular utility in treating patients with central sleep apnea or Hunter–Cheyne–Stokes breathing (CSA/HCSB) as well as combinations of CSA/HCSB and obstructive sleep apnea (OSA), and their use is the subject of this review.
significant amount of airflow that escapes from the mask both on purpose (through ports in or near the mask to wash out exhaled CO2) and by accident (mask or mouth leak). The method by which this is accomplished entails the continuous measurement of delivered airflow and system pressure along with calculations (performed by a microprocessor) that can best be described as deriving an approximation to actual nasal/oral airflow. The ResMed flow generator likely utilizes the series of calculations shown in Fig. 1a [3,4], and the approach used by the Respironics apparatus appears to be one of the two processes in Fig. 1b [5,6]. In both devices, the operator may set a variety of fixed or automatically calculated back-up rates and both now also include the ability to auto-titrate EPAP to maintain airway patency. The latter algorithms are well beyond the scope of this discussion.
&
REVIEW OF ADAPTIVE SERVOVENTILATION TECHNOLOGY The technology behind ASV is actually quite complicated and beyond the scope of this discussion, but a brief overview (which will be limited to the two devices available in both the United States and overseas) is warranted in order to provide an appreciation of the assumptions and approximations implemented in these devices. This technology, as much as it is available from examination of applicable patents and marketing literature (as precise details are proprietary and not disclosed), has been reviewed in detail elsewhere [1 ,2 ]. In essence, ASV is a negative feedback control system in which a measurement of actual minute ventilation (ResMed) or peak inspiratory airflow (Respironics) is compared with a target value and the difference (error) is used to vary IPAP (and therefore pressure support) on a breath-by-breath basis. The target value for the ResMed flow generator is a percentage (90–95%) of a weighted average of minute ventilation from the preceding 9 min or so with more recent breaths reflected more heavily in the average. The Respironics apparatus takes a similar approach, only targeting the average of peak inspiratory flow and using a moving window of about 4 min. The difficulty inherent in this technology lies in measuring instantaneous airflow without a direct connection to the airway, and despite the presence of the &
&
HOW DOES ADAPTIVE SERVOVENTILATION SUPPRESS CENTRAL SLEEP APNEA/HUNTER–CHEYNE–STOKES BREATHING? The theoretical basis underlying the ability of ASV to suppress CSA/HCSB requires that the time constant, controller gain, and other attributes of the feedback algorithm be chosen so as to ensure that the amount of pressure support supplied by the device is anticyclic to the patient’s own ventilation as a reflection of ventilatory drive [1 ]. ASV was developed as a treatment modality for hypocapnic periodic breathing, that is, CSA/HCSB. The pathogenesis of CSA/ HCSB is largely explained by the presence of an apneic threshold for PaCO2 that exists only during non rapid eye movement sleep, when the supratentorial contribution to ventilatory drive is lost [7–9], whereas the SDB associated with opioids or neuromuscular ventilatory failure is generally accompanied by varying degrees of hypercapnia and is frequently less amenable to ASV therapy. CSA/HCSB is characterized by central apneas or hypopneas interposed between intervals of hyperventilation; HCSB is differentiated from CSA by a crescendo-decrescendo pattern of tidal volumes during the intervals of hyperventilation. The central hypopneas and apneas increase PaCO2, stimulating chemoreceptors and prompting the ensuing period of hyperventilation. The hyperventilation depresses PaCO2 to a level at or below the apneic threshold, resulting in decrescendo breathing followed by a central hypopnea or apnea (in HCSB) or an abrupt central apnea in CSA. The time that passes before changes in pulmonary venous pCO2 are detected by the central chemoreceptor, and excessive controller gain are factors thought to destabilize ventilatory
1070-5287 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
&
www.co-pulmonarymedicine.com
551
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Sleep and respiratory neurobiology
(a) Compute leakage conductance (inverse of resistance) from instantaneous system airflow divided by the square root of instantaneous system pressure
Compute mask leak from leakage conductance multiplied by the square root of instantaneous system pressure
Subtract mask leak from total airflow delivered to mask to determine instantaneous respiratory airflow
(b) Compute average mask leak from instantaneous airflow averaged over multiple full breaths, thus averaging out inspiratory and expiratory tidal volumes
Calculate average leak pathway conductance for a complete breath, as in VPAP adapt
Instantaneous respiratory airflow computed by subtracting average leak from total instantaneous mask airflow
Compute instantaneous mask leak airflow by multiplying average leak conductance by square root of instantaneous circuit pressure
Subtract instantaneous mask leak airflow from instantaneous mask airflow to derive instantaneous respiratory airflow
FIGURE 1. (a) Method appearing in applicable patent for approximating breath-by-breath instantaneous airflow most likely utilized in the ResMed S9 VPAP Adapt. (b) The two methods appearing in applicable patents for approximating breath-bybreath instantaneous airflow that are most likely to be utilized in the Philips Respironics bilevel positive airway pressure autoSV Advanced – System One.
control, and the periodic breathing of CSA/HCSB is a direct manifestation of cyclic variations in inspiratory drive [10]. A similar phenomenon likely is responsible for CSA/HCSB in patients with cerebrovascular disease [11]. ASV counterbalances this waxing and waning of inspiratory drive by varying the degree of pressure support: decreasing pressure support when inspiratory drive is high and increasing pressure support when inspiratory drive declines. Consequently, the combination of the patient’s own inspiratory drive and that of the ASV device sums to maintain a more constant degree of ventilatory drive, thus damping the variations in tidal volume [12,13]. 552
www.co-pulmonarymedicine.com
RECENT ADVANCES The salutary effects of the acute application of ASV to patients with CSA/HCSB due to congestive heart failure (CHF) can no longer be disputed, as well as the overall superiority in this regard for ASV over CPAP, biPAP-S, biPAP-S/T, or oxygen monotherapy [14,15 ]. In this setting, polysomnography or respiratory polygraphy consistently shows a higher degree of improvement in, and more patients with normalized values of, apnea-hypopnea index (AHI), oxyhemoglobin desaturation index, and parameters related to sleep continuity. The last 18 months have seen increasing publication of short-term and longterm studies examining surrogate endpoints related &
Volume 20 Number 6 November 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Adaptive servo-ventilation Brown and Javaheri
to CHF morbidity including randomized controlled trials (RCTs). In addition, studies related to technical aspects of ASV have also appeared.
Short-term results of adaptive servoventilation treatment for central sleep apnea/ hunter–cheyne–stokes breathing in congestive heart failure Javaheri et al. [14] have called attention to the possibility that injudicious choice of ASV settings, particularly of EPAP with a resulting increase in intrathoracic pressure, might have a deleterious effect on cardiac hemodynamics due to excessive reductions in preload. Conversely, the use of ASV could actually improve hemodynamics due to the reduction in preload in combination with reductions in afterload. Yamada et al. [16 ] examined this issue during right heart catheterization in 11 normal controls and 34 patients with chronic, stable CHF. Patients had ejection fraction by echocardiography of greater than 45%, New York Heart Association (NYHA) functional class II or III, and plasma brain natriuretic peptide (BNP) greater than 100 pg/ml. Control patients were suspected to have ischemic heart disease, but no significant coronary artery disease following coronary angiography. None of the patients had SDB. Measurements were made before and after 15 min of ASV treatment using a ResMed Autoset CS flow generator (a predecessor of the S9 VPAP Adapt) using the default settings of EPAP ¼ 5 cm H2O and pressure support range of 3–10 H2O; back-up rate setting is not reported but was probably the default value. Outcome measures were echocardiographic variables, BNP, and hemodynamic variables from the right heart catheterization. Stroke volume index (SVI), the primary outcome measure, declined slightly in the control group exposed to ASV, as might be expected in individuals with normal hearts operating on the Frank–Starling curve. Individual patients experienced either increased SVI (15/34) or decreased SVI (19/34). Heart rate, systolic blood pressure, and pulmonary capillary wedge pressure (PCWP) did not change in either group. After multivariate analysis, significant predictors of improvement in SVI were baseline PCWP and mitral regurgitation area/left atrial area, which revealed a linear relationship. The authors speculated that the latter finding was related to improvement in left ventricular size because of reduced preload; left ventricular dilatation is known to produce functional mitral regurgitation by stretching the papillary muscles, thereby tethering the mitral valve leaflets and preventing full closure. Unfortunately, the authors did not include hemodynamic and echocardiographic assessments &&
that could have supported this mechanism. Nevertheless, other investigators have reported that the application of CPAP or biPAP acutely reduces functional mitral regurgitation (when increasing ejection fraction) during exacerbations of CHF [17]. The results reported by Yamada et al. were limited by the relatively low values of EPAP utilized and the lack of detailed information concerning the level of pressure support being applied (a minimum of 3 cm H2O and a maximum of 10 cm H2O, which translates to settings as high as 15/5 cm H2O or as low as 8/5 cm H2O depending on whether periodic breathing (PB) was or was not present). The authors reported that the patients were awake, but there was no electroencephalographic documentation of sleep/wake status; as patients are usually sedated for cardiac catheterization, some of the patients may have been asleep and exhibiting greater degrees of PB, although PB can occasionally be present in CHF patients even when awake. Adverse outcomes in CHF have been attributed, in part, to excessive beta-1 adrenergic tone; hence the use of central nervous system-active betaadrenergic antagonists has become a standard of care [18]. Patients with CHF complicated by CSA/ HCSB are known to express additional sympathetic over activity, and therefore reductions in sympathetic tone resulting from ASV treatment of CSA/HCSB could provide further support for using this device in such patients. An early study was reported by Pepperell et al. [19] in patients with chronic symptomatic systolic heart failure (NYHA class II–IV) and SDB that was almost exclusively CSA/HCSB. Using a prospective RCT design comparing Autoset CS (default settings) with sham ASV, these investigators found that urinary metanephrines declined after 4 weeks of ASV treatment while increasing slightly after sham ASV. More recently, Ushijima et al. [20] examined this issue in 57 patients, 70% of whom had CSA/HCSB of varying degrees without OSA, left ventricular ejection fraction (LVEF) less than 45% and NYHA functional class I, II, or III. All patients underwent acclimatization to ASV for 30 min. Patients were randomized between ASV (n ¼ 29; ResMed Autoset CS at default settings for 20/29 and EPAP ¼ 4 cm H2O/maximum pressure support ¼ 8 cm H2O for 9/29) and CPAP (n ¼ 28; using the median airway pressure obtained during the acclimatization). A total of 76 and 71% of the patients in each group, respectively, were receiving chronic beta-blocker therapy. Integrated skeletal muscle sympathetic nerve activity (SNA), a reflection of central nervous system sympathetic outflow, was measured during 10 min without treatment and then during 30 min of either ASV or CPAP application. The degree of muscle SNA declined
1070-5287 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.co-pulmonarymedicine.com
553
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Sleep and respiratory neurobiology
significantly from baseline in the ASV group but not in the CPAP group. Unfortunately, these investigators did not include polysomnographic (PSG) monitoring as part of the experimental protocol, and therefore it is not known what the effect of either modality might have been on SDB, nor is it definitively known whether the patients were awake or asleep. It is known that CSA/HCSB is more prevalent in patients with chronic kidney disease (CKD), and many of these patients also suffer from CHF. The severity of SDB seems to correlate with both CKD stage and the severity of CHF, when studied retrospectively [21]. However, whether there is a bidirectional effect (as discussed by Jhamb and Unruh [22] for OSA in this issue) in which treatment of CSA/ HCSB affects CKD outcome remains unknown. A study by Yoshihisa et al. [23 ] recently investigated this issue in 50 patients with stable CHF (NYHA class >II, LVEF 5 h/day) revealed significant improvements in left ventricular endsystolic volume and LVEF. Interestingly, there was also improvement in tricuspid annular plane systolic excursion, a measure of right ventricular function. In terms of performance status, at baseline the CSA/HCSB patients were distributed between NYHA class II (n ¼ 5) and class III (n ¼ 4). After 6 months of ASV use, patients transitioned toward lower NYHA class, including class I (n ¼ 2), class II (n ¼ 6), and class III (n ¼ 2). Similar results were reported by Koyama et al. [25] in a study of 19 patients with CHF with predominantly CSA/HCSB on PSG. They studied 10 patients who were compliant with ASV treatment (6.4 0.6 h/night for 6 months using an Autoset CS with default programming) compared with nine who refused this therapy or exhibited minimal use (0.4 0.5 h/night) and remained on conventional pharmacological management. There were no improvements in echocardiographic metrics of left ventricular function or of plasma BNP levels in the latter group, whereas those who used ASV exhibited improvements in LVEF, left ventricular end-diastolic volume, left ventricular end-systolic volume, and BNP. These investigators also assessed cardiac SNA using 123meta-iodobenzylguanidine myocardial scintigraphy. Both washout ratio and delayed heart-to-mediastinal ratio were used as measures of cardiac SNA. At baseline, washout ratio and heart-to-mediastinal ratio correlated with AHI and central apnea index, indicating that CSA/HCSB was associated with greater cardiac SNA. Both indices improved in the ASV group but not in those patients with no or minimal ASV adherence after 6 months of follow-up. ASV may have a role in ameliorating renal failure in patients with CSA/HCSB and CHF. Owada et al. [26 ] performed an observational study in 80 patients with stable CHF (NYHA class >II), CKD (eGFR 15/h). Patients who accepted and were adherent to ASV use (n ¼ 36) were compared with those who refused ASV or discontinued its use (n ¼ 44). After 6 months, the ASV group demonstrated improvements in NYHA class, BNP, creatinine, cystatin C, noradrenaline, and eGFR as well as the echocardiographic &&
Volume 20 Number 6 November 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Adaptive servo-ventilation Brown and Javaheri
measurements of LVEF, left ventricular mass index, left atrial volume index, and E/E’. None of these metrics improved in the non-ASV group. Lastly, a robust improvement in event-free survival was enjoyed by those individuals using ASV at home. Four recent studies are notable for their use of an RCT study design, one of which included outcome variables that measured prognosis. Yoshihisa et al. [27 ] randomized 36 patients with heart failure and preserved ejection fraction (diastolic dysfunction) and moderate-to-severe, predominantly, central SDB to receive treatment with either ASV plus conventional pharmacotherapy or pharmacotherapy alone. Titration of ASV was performed in a manner similar to that described by these investigators in an earlier study, detailed above [23 ]. In addition to obtaining echocardiography, plasma BNP and highsensitivity troponin T, and eGFR (Modification of Diet in Renal Disease formula), they also analyzed event-free survival, in which events were defined as cardiac death or rehospitalization. After 6 months of follow-up, NYHA class, BNP, and two measures of diastolic function [ratio of early transmitral flow velocity to mitral annular velocity (E/E’) and left atrial volume index] improved in the ASV group, whereas high-sensitivity troponin-T was unchanged. Of particular interest given the negative results of the CanPAP trial, [28] event-free survival was appreciably higher in the ASV-treated individuals compared with those receiving pharmacotherapy only. The second RCT was reported by Kasai et al. [29 ] and examined neuro-hormonal measures after 3 months of ASV treatment in patients with stable systolic CHF (LVEF
Adaptive servo-ventilation for the treatment of central sleep apnea in congestive heart failure: what have we learned? Lee K. Brown a and Shahrokh Javaheri b
Purpose of review Positive airway pressure devices for the noninvasive treatment of sleep-disordered breathing are being marketed that have substantially expanded capabilities. Most recently, adaptive servo-ventilation devices have become available that are capable of measuring patient ventilation continuously and use that information to adjust expiratory positive airway pressure and pressure support levels to abolish central and obstructive apneas and hypopneas, including central sleep-disordered breathing of the Hunter–Cheyne– Stokes variety. Patients with congestive heart failure are particularly prone to developing central sleep apnea and/or Hunter–Cheyne–Stokes breathing, and studies have shown that suppression of these abnormal breathing patterns may improve cardiac function and, ultimately, mortality. Recent findings Over the last approximately 18 months, increasing numbers of studies have appeared demonstrating improvement in cardiac function and other important outcomes after both acute application of adaptive servo-ventilation as well as 3 to 6 months of use in patients with congestive heart failure and central sleep apnea/Hunter–Cheyne–Stokes breathing. Several of these studies are randomized controlled trials and several include assessment of cardiac event-free survival showing an advantage to treating with adaptive servo-ventilation. Summary As an adjunct to optimal pharmacological management, adaptive servo-ventilation shows considerable promise as a means to improve outcomes in patients with congestive heart failure complicated by central sleep apnea/Hunter–Cheyne–Stokes breathing. Larger randomized controlled trials will be necessary to demonstrate the ultimate role of this therapeutic modality in such patients. Keywords adaptive servo-ventilation, central sleep apnea, congestive heart failure, Hunter–Cheyne–Stokes breathing
INTRODUCTION The last decade has seen the introduction of evermore technologically advanced devices for the noninvasive treatment of sleep-disordered breathing (SDB). Upon the original, straightforward platform of the continuous positive airway pressure (CPAP) flow generator, first introduced in 1981, have been built devices with a dizzying array of capabilities. These include bilevel positive airway pressure (biPAP) flow generators that switch between a lower pressure during expiration [expiratory positive airway pressure (EPAP)] and a higher pressure during inspiration [inspiratory positive airway pressure (IPAP)]; automatically titrating flow generators that detect decreases in airflow and signs of upper airway obstruction in order to vary CPAP or biPAP settings; and devices that target either a set level of www.co-pulmonarymedicine.com
ventilation (average volume assured pressure support) or a proportion of the patient’s native ventilation or inspiratory flow [adaptive servoventilation (ASV)]. The latter devices now combine their original technology with auto-titrating EPAP. There are currently two manufacturers marketing ASV flow generators in the United States (ResMed, a Department of Internal Medicine, University of New Mexico School of Medicine, Albuquerque, New Mexico, bUniversity of Cincinnati College of Medicine, and Cincinnati, Ohio, USA
Correspondence to Lee K. Brown, MD, Program in Sleep Medicine, University of New Mexico Health Sciences Center, 1101 Medical Arts Avenue NE, Building 2, Albuquerque, NM 87102, USA. Tel: +1 505 272 6110; fax: +1 505 272 6112; e-mail: [email protected] Curr Opin Pulm Med 2014, 20:550–557 DOI:10.1097/MCP.0000000000000108 Volume 20 Number 6 November 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Adaptive servo-ventilation Brown and Javaheri
KEY POINTS ASV, when properly titrated, has proven efficacy in suppressing CSA/HCSB in patients with CHF. Recent investigations, including several RCTs of high quality, have shown improvement in cardiac function, exercise capacity, and event-free survival when CHF patients receive ASV therapy to suppress CSA/HCSB. Larger RCTs will be necessary to confirm the proper role of ASV as a component of CHF management.
San Diego, California markets the S9 VPAP Adapt, and Philips Respironics, Murrysville, Pennsylvania markets the biPAP autoSV Advanced – System One) and a third device is available outside the United ¨r Medizin GmbH þ Co., States (Weinmann Gera¨te fu Hamburg, Germany produces the SOMNOvent CR) [1 ]. All three devices have shown particular utility in treating patients with central sleep apnea or Hunter–Cheyne–Stokes breathing (CSA/HCSB) as well as combinations of CSA/HCSB and obstructive sleep apnea (OSA), and their use is the subject of this review.
significant amount of airflow that escapes from the mask both on purpose (through ports in or near the mask to wash out exhaled CO2) and by accident (mask or mouth leak). The method by which this is accomplished entails the continuous measurement of delivered airflow and system pressure along with calculations (performed by a microprocessor) that can best be described as deriving an approximation to actual nasal/oral airflow. The ResMed flow generator likely utilizes the series of calculations shown in Fig. 1a [3,4], and the approach used by the Respironics apparatus appears to be one of the two processes in Fig. 1b [5,6]. In both devices, the operator may set a variety of fixed or automatically calculated back-up rates and both now also include the ability to auto-titrate EPAP to maintain airway patency. The latter algorithms are well beyond the scope of this discussion.
&
REVIEW OF ADAPTIVE SERVOVENTILATION TECHNOLOGY The technology behind ASV is actually quite complicated and beyond the scope of this discussion, but a brief overview (which will be limited to the two devices available in both the United States and overseas) is warranted in order to provide an appreciation of the assumptions and approximations implemented in these devices. This technology, as much as it is available from examination of applicable patents and marketing literature (as precise details are proprietary and not disclosed), has been reviewed in detail elsewhere [1 ,2 ]. In essence, ASV is a negative feedback control system in which a measurement of actual minute ventilation (ResMed) or peak inspiratory airflow (Respironics) is compared with a target value and the difference (error) is used to vary IPAP (and therefore pressure support) on a breath-by-breath basis. The target value for the ResMed flow generator is a percentage (90–95%) of a weighted average of minute ventilation from the preceding 9 min or so with more recent breaths reflected more heavily in the average. The Respironics apparatus takes a similar approach, only targeting the average of peak inspiratory flow and using a moving window of about 4 min. The difficulty inherent in this technology lies in measuring instantaneous airflow without a direct connection to the airway, and despite the presence of the &
&
HOW DOES ADAPTIVE SERVOVENTILATION SUPPRESS CENTRAL SLEEP APNEA/HUNTER–CHEYNE–STOKES BREATHING? The theoretical basis underlying the ability of ASV to suppress CSA/HCSB requires that the time constant, controller gain, and other attributes of the feedback algorithm be chosen so as to ensure that the amount of pressure support supplied by the device is anticyclic to the patient’s own ventilation as a reflection of ventilatory drive [1 ]. ASV was developed as a treatment modality for hypocapnic periodic breathing, that is, CSA/HCSB. The pathogenesis of CSA/ HCSB is largely explained by the presence of an apneic threshold for PaCO2 that exists only during non rapid eye movement sleep, when the supratentorial contribution to ventilatory drive is lost [7–9], whereas the SDB associated with opioids or neuromuscular ventilatory failure is generally accompanied by varying degrees of hypercapnia and is frequently less amenable to ASV therapy. CSA/HCSB is characterized by central apneas or hypopneas interposed between intervals of hyperventilation; HCSB is differentiated from CSA by a crescendo-decrescendo pattern of tidal volumes during the intervals of hyperventilation. The central hypopneas and apneas increase PaCO2, stimulating chemoreceptors and prompting the ensuing period of hyperventilation. The hyperventilation depresses PaCO2 to a level at or below the apneic threshold, resulting in decrescendo breathing followed by a central hypopnea or apnea (in HCSB) or an abrupt central apnea in CSA. The time that passes before changes in pulmonary venous pCO2 are detected by the central chemoreceptor, and excessive controller gain are factors thought to destabilize ventilatory
1070-5287 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
&
www.co-pulmonarymedicine.com
551
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Sleep and respiratory neurobiology
(a) Compute leakage conductance (inverse of resistance) from instantaneous system airflow divided by the square root of instantaneous system pressure
Compute mask leak from leakage conductance multiplied by the square root of instantaneous system pressure
Subtract mask leak from total airflow delivered to mask to determine instantaneous respiratory airflow
(b) Compute average mask leak from instantaneous airflow averaged over multiple full breaths, thus averaging out inspiratory and expiratory tidal volumes
Calculate average leak pathway conductance for a complete breath, as in VPAP adapt
Instantaneous respiratory airflow computed by subtracting average leak from total instantaneous mask airflow
Compute instantaneous mask leak airflow by multiplying average leak conductance by square root of instantaneous circuit pressure
Subtract instantaneous mask leak airflow from instantaneous mask airflow to derive instantaneous respiratory airflow
FIGURE 1. (a) Method appearing in applicable patent for approximating breath-by-breath instantaneous airflow most likely utilized in the ResMed S9 VPAP Adapt. (b) The two methods appearing in applicable patents for approximating breath-bybreath instantaneous airflow that are most likely to be utilized in the Philips Respironics bilevel positive airway pressure autoSV Advanced – System One.
control, and the periodic breathing of CSA/HCSB is a direct manifestation of cyclic variations in inspiratory drive [10]. A similar phenomenon likely is responsible for CSA/HCSB in patients with cerebrovascular disease [11]. ASV counterbalances this waxing and waning of inspiratory drive by varying the degree of pressure support: decreasing pressure support when inspiratory drive is high and increasing pressure support when inspiratory drive declines. Consequently, the combination of the patient’s own inspiratory drive and that of the ASV device sums to maintain a more constant degree of ventilatory drive, thus damping the variations in tidal volume [12,13]. 552
www.co-pulmonarymedicine.com
RECENT ADVANCES The salutary effects of the acute application of ASV to patients with CSA/HCSB due to congestive heart failure (CHF) can no longer be disputed, as well as the overall superiority in this regard for ASV over CPAP, biPAP-S, biPAP-S/T, or oxygen monotherapy [14,15 ]. In this setting, polysomnography or respiratory polygraphy consistently shows a higher degree of improvement in, and more patients with normalized values of, apnea-hypopnea index (AHI), oxyhemoglobin desaturation index, and parameters related to sleep continuity. The last 18 months have seen increasing publication of short-term and longterm studies examining surrogate endpoints related &
Volume 20 Number 6 November 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Adaptive servo-ventilation Brown and Javaheri
to CHF morbidity including randomized controlled trials (RCTs). In addition, studies related to technical aspects of ASV have also appeared.
Short-term results of adaptive servoventilation treatment for central sleep apnea/ hunter–cheyne–stokes breathing in congestive heart failure Javaheri et al. [14] have called attention to the possibility that injudicious choice of ASV settings, particularly of EPAP with a resulting increase in intrathoracic pressure, might have a deleterious effect on cardiac hemodynamics due to excessive reductions in preload. Conversely, the use of ASV could actually improve hemodynamics due to the reduction in preload in combination with reductions in afterload. Yamada et al. [16 ] examined this issue during right heart catheterization in 11 normal controls and 34 patients with chronic, stable CHF. Patients had ejection fraction by echocardiography of greater than 45%, New York Heart Association (NYHA) functional class II or III, and plasma brain natriuretic peptide (BNP) greater than 100 pg/ml. Control patients were suspected to have ischemic heart disease, but no significant coronary artery disease following coronary angiography. None of the patients had SDB. Measurements were made before and after 15 min of ASV treatment using a ResMed Autoset CS flow generator (a predecessor of the S9 VPAP Adapt) using the default settings of EPAP ¼ 5 cm H2O and pressure support range of 3–10 H2O; back-up rate setting is not reported but was probably the default value. Outcome measures were echocardiographic variables, BNP, and hemodynamic variables from the right heart catheterization. Stroke volume index (SVI), the primary outcome measure, declined slightly in the control group exposed to ASV, as might be expected in individuals with normal hearts operating on the Frank–Starling curve. Individual patients experienced either increased SVI (15/34) or decreased SVI (19/34). Heart rate, systolic blood pressure, and pulmonary capillary wedge pressure (PCWP) did not change in either group. After multivariate analysis, significant predictors of improvement in SVI were baseline PCWP and mitral regurgitation area/left atrial area, which revealed a linear relationship. The authors speculated that the latter finding was related to improvement in left ventricular size because of reduced preload; left ventricular dilatation is known to produce functional mitral regurgitation by stretching the papillary muscles, thereby tethering the mitral valve leaflets and preventing full closure. Unfortunately, the authors did not include hemodynamic and echocardiographic assessments &&
that could have supported this mechanism. Nevertheless, other investigators have reported that the application of CPAP or biPAP acutely reduces functional mitral regurgitation (when increasing ejection fraction) during exacerbations of CHF [17]. The results reported by Yamada et al. were limited by the relatively low values of EPAP utilized and the lack of detailed information concerning the level of pressure support being applied (a minimum of 3 cm H2O and a maximum of 10 cm H2O, which translates to settings as high as 15/5 cm H2O or as low as 8/5 cm H2O depending on whether periodic breathing (PB) was or was not present). The authors reported that the patients were awake, but there was no electroencephalographic documentation of sleep/wake status; as patients are usually sedated for cardiac catheterization, some of the patients may have been asleep and exhibiting greater degrees of PB, although PB can occasionally be present in CHF patients even when awake. Adverse outcomes in CHF have been attributed, in part, to excessive beta-1 adrenergic tone; hence the use of central nervous system-active betaadrenergic antagonists has become a standard of care [18]. Patients with CHF complicated by CSA/ HCSB are known to express additional sympathetic over activity, and therefore reductions in sympathetic tone resulting from ASV treatment of CSA/HCSB could provide further support for using this device in such patients. An early study was reported by Pepperell et al. [19] in patients with chronic symptomatic systolic heart failure (NYHA class II–IV) and SDB that was almost exclusively CSA/HCSB. Using a prospective RCT design comparing Autoset CS (default settings) with sham ASV, these investigators found that urinary metanephrines declined after 4 weeks of ASV treatment while increasing slightly after sham ASV. More recently, Ushijima et al. [20] examined this issue in 57 patients, 70% of whom had CSA/HCSB of varying degrees without OSA, left ventricular ejection fraction (LVEF) less than 45% and NYHA functional class I, II, or III. All patients underwent acclimatization to ASV for 30 min. Patients were randomized between ASV (n ¼ 29; ResMed Autoset CS at default settings for 20/29 and EPAP ¼ 4 cm H2O/maximum pressure support ¼ 8 cm H2O for 9/29) and CPAP (n ¼ 28; using the median airway pressure obtained during the acclimatization). A total of 76 and 71% of the patients in each group, respectively, were receiving chronic beta-blocker therapy. Integrated skeletal muscle sympathetic nerve activity (SNA), a reflection of central nervous system sympathetic outflow, was measured during 10 min without treatment and then during 30 min of either ASV or CPAP application. The degree of muscle SNA declined
1070-5287 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.co-pulmonarymedicine.com
553
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Sleep and respiratory neurobiology
significantly from baseline in the ASV group but not in the CPAP group. Unfortunately, these investigators did not include polysomnographic (PSG) monitoring as part of the experimental protocol, and therefore it is not known what the effect of either modality might have been on SDB, nor is it definitively known whether the patients were awake or asleep. It is known that CSA/HCSB is more prevalent in patients with chronic kidney disease (CKD), and many of these patients also suffer from CHF. The severity of SDB seems to correlate with both CKD stage and the severity of CHF, when studied retrospectively [21]. However, whether there is a bidirectional effect (as discussed by Jhamb and Unruh [22] for OSA in this issue) in which treatment of CSA/ HCSB affects CKD outcome remains unknown. A study by Yoshihisa et al. [23 ] recently investigated this issue in 50 patients with stable CHF (NYHA class >II, LVEF 5 h/day) revealed significant improvements in left ventricular endsystolic volume and LVEF. Interestingly, there was also improvement in tricuspid annular plane systolic excursion, a measure of right ventricular function. In terms of performance status, at baseline the CSA/HCSB patients were distributed between NYHA class II (n ¼ 5) and class III (n ¼ 4). After 6 months of ASV use, patients transitioned toward lower NYHA class, including class I (n ¼ 2), class II (n ¼ 6), and class III (n ¼ 2). Similar results were reported by Koyama et al. [25] in a study of 19 patients with CHF with predominantly CSA/HCSB on PSG. They studied 10 patients who were compliant with ASV treatment (6.4 0.6 h/night for 6 months using an Autoset CS with default programming) compared with nine who refused this therapy or exhibited minimal use (0.4 0.5 h/night) and remained on conventional pharmacological management. There were no improvements in echocardiographic metrics of left ventricular function or of plasma BNP levels in the latter group, whereas those who used ASV exhibited improvements in LVEF, left ventricular end-diastolic volume, left ventricular end-systolic volume, and BNP. These investigators also assessed cardiac SNA using 123meta-iodobenzylguanidine myocardial scintigraphy. Both washout ratio and delayed heart-to-mediastinal ratio were used as measures of cardiac SNA. At baseline, washout ratio and heart-to-mediastinal ratio correlated with AHI and central apnea index, indicating that CSA/HCSB was associated with greater cardiac SNA. Both indices improved in the ASV group but not in those patients with no or minimal ASV adherence after 6 months of follow-up. ASV may have a role in ameliorating renal failure in patients with CSA/HCSB and CHF. Owada et al. [26 ] performed an observational study in 80 patients with stable CHF (NYHA class >II), CKD (eGFR 15/h). Patients who accepted and were adherent to ASV use (n ¼ 36) were compared with those who refused ASV or discontinued its use (n ¼ 44). After 6 months, the ASV group demonstrated improvements in NYHA class, BNP, creatinine, cystatin C, noradrenaline, and eGFR as well as the echocardiographic &&
Volume 20 Number 6 November 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Adaptive servo-ventilation Brown and Javaheri
measurements of LVEF, left ventricular mass index, left atrial volume index, and E/E’. None of these metrics improved in the non-ASV group. Lastly, a robust improvement in event-free survival was enjoyed by those individuals using ASV at home. Four recent studies are notable for their use of an RCT study design, one of which included outcome variables that measured prognosis. Yoshihisa et al. [27 ] randomized 36 patients with heart failure and preserved ejection fraction (diastolic dysfunction) and moderate-to-severe, predominantly, central SDB to receive treatment with either ASV plus conventional pharmacotherapy or pharmacotherapy alone. Titration of ASV was performed in a manner similar to that described by these investigators in an earlier study, detailed above [23 ]. In addition to obtaining echocardiography, plasma BNP and highsensitivity troponin T, and eGFR (Modification of Diet in Renal Disease formula), they also analyzed event-free survival, in which events were defined as cardiac death or rehospitalization. After 6 months of follow-up, NYHA class, BNP, and two measures of diastolic function [ratio of early transmitral flow velocity to mitral annular velocity (E/E’) and left atrial volume index] improved in the ASV group, whereas high-sensitivity troponin-T was unchanged. Of particular interest given the negative results of the CanPAP trial, [28] event-free survival was appreciably higher in the ASV-treated individuals compared with those receiving pharmacotherapy only. The second RCT was reported by Kasai et al. [29 ] and examined neuro-hormonal measures after 3 months of ASV treatment in patients with stable systolic CHF (LVEF
